A general model for deriving preference structures from data

In this paper we comment upon the integrated model for valued preferences introduced by Fodor, Ovchinnikov and Roubens. In particular, while on one hand we revise basic assumptions and point out their intuitive meaning, on the other hand we propose an alternative mathematical justification of such a model which allows not only a better understanding of the obtained results, but also a functional characterization of the whole family of solutions. (C) 1997 Elsevier Science B.V.